Efficiency of adaptive cluster sampling and traditional sampling for coastal mangrove in Hainan of China

Abstract

Based on two species of Coastal Mangrove in Hainan of China, Sonneratia Apetala Buch-Ham and Sonneratia caseoli, we estimated the density of the two species to evaluate the efficiency of adaptive cluster sampling (ACS), simple random sampling (SRS) and traditional systematic sampling (SYS). Our initial experimental designs for ACS consisted of 5 unit areas, 6 initial sampling proportions, 4 initial sample sizes and 5 criterion values in 1,000 repetitions. From the aspect of factors influencing efficiency, we analysed the efficiency of ACS in various designs. We also compared the efficiencies of the three methods on the indexes of the relative error, the variance of density estimator and the relative sampling efficiencies. We found that ACS yielded smaller variance than the traditional sampling methods. ACS was a powerful sampling method when a population was spatially aggregated. We also determined the optimum unit area for the two species studied using the two estimators (HT and HH) of adaptive cluster sampling. They were 20 m<sup>2 </sup>(2 × 10 m), 15 m<sup>2 </sup>(3 × 5 m) for S. Apetala Buch-Ham and 25 m<sup>2 </sup>(5 × 5 m), 15 m<sup>2 </sup>(3 × 5 m) for S. caseolari, respectively.